Big data analytics for improving customer win-back rate in townhome segment

Abstract

Thailand’s real-estate market is now facing with negative growth, where most developers are encountered with challenges in generating satisfactory revenue, particularly from new customers. In order to improve the revenue stream, customer win-back approach that focuses on current customers in the sales funnel and re-engages them by offering more attractive residential projects is therefore initiated. In particular, we focus on the development of predictive analytics model that potentially increases customer win-back rate by exploring the current databases, identifying significance for win-back (based on past performance), and testing the resulting predictions via several machine learning algorithms. The proposed method returns a propensity model that ranks customers based on likelihood to purchase if contacted. We test the proposed method by running an A/B test for 3 weeks and then compare (1) average number of refer cases to sales department and (2) average customer referral rate with the base case. We find that the propensity model impressively helps increase average number of refer cases to sales department by 11.8% and increase referral rate by 13.5%. In terms of finance, these significant improvements lead to a revenue uplift of 20.9% year-on-year, valued at THB 253.5 million.